Ontology drift is a challenge for explainable data governance

Regulatory compliance increasingly requires explainable AI: The Basel Committee's Standard No. 239 mandates that financial institutions use explainable AI to demonstrate data quality and produce appropriate risk reports for different stakeholders, making transparency a legal requirement rather than optional.

Data taxonomies must evolve continuously: Financial institutions face a specific compliance challenge in maintaining complete, firm-wide data taxonomies that accurately reflect their operations. The paper identifies "ontology drift"—the constant evolution of how financial data is categorized and understood—as a persistent obstacle to staying compliant.

Static AI systems cannot meet dynamic regulatory needs: Because financial ontologies continuously change, one-time AI implementations cannot ensure ongoing compliance. Organizations must build systems designed for continuous updating rather than treating compliance as a one-time project.

Explainability serves multiple stakeholders with different needs: The paper emphasizes that explainable AI in finance must communicate effectively to various audiences (regulators, internal risk teams, executives), requiring governance systems that can adapt explanations rather than provide single standardized outputs.

Data governance is foundational to responsible AI in finance: Before deploying AI for decision-making or risk assessment, financial institutions must first solve the underlying data governance problem—ensuring their data definitions and taxonomies are clear, documented, and kept current.